EEG denoising with a recurrent quantum neural network for a brain-computer interface

V Gandhi, V Arora, Laxmidhar Behera, G Prasad, DH Coyle, TM McGinnity

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

Brain-computer interface (BCI) technology is a means of communication that allows individuals with severe movement disability to communicate with external assistive devices using the electroencephalogram (EEG) or other brain signals. This paper presents an alternative neural information processing architecture using the Schrödinger wave equation (SWE) for enhancement of the raw EEG signal. The raw EEG signal obtained during the motor imagery (MI) of a BCI user is intrinsically embedded with non-Gaussian noise while the actual signal is still a mystery. The proposed work in the field of recurrent quantum neural network (RQNN) is designed to filter such non-Gaussian noise using an unsupervised learning scheme without making any assumption about the signal type. The proposed learning architecture has been modified to do away with the Hebbian learning associated with the existing RQNN architecture as this learning scheme was found to be unstable for complex signals such as EEG. Besides, this the soliton behaviour of the non-linear SWE was not properly preserved in the existing scheme. The unsupervised learning algorithm proposed in this paper is able to efficiently capture the statistical behaviour of the input signal while making the algorithm robust to parametric sensitivity. This denoised EEG signal is then fed as an input to the feature extractor to obtain the Hjorth features. These features are then used to train a Linear Discriminant Analysis (LDA) classifier. It is shown that the accuracy of the classifier output over the training and the evaluation datasets using the filtered EEG is much higher compared to that using the raw EEG signal. The improvement in classification accuracy computed over nine subjects is found to be statistically significant.
LanguageEnglish
Title of host publicationUnknown Host Publication
Pages1583-1590
Number of pages8
DOIs
Publication statusPublished - 31 Jul 2011
EventThe 2011 International Joint Conference on Neural Networks - San Jose, CA
Duration: 31 Jul 2011 → …

Conference

ConferenceThe 2011 International Joint Conference on Neural Networks
Period31/07/11 → …

Fingerprint

Brain computer interface
Electroencephalography
Neural networks
Unsupervised learning
Wave equations
Classifiers
Discriminant analysis
Solitons
Network architecture
Learning algorithms
Brain
Communication

Cite this

Gandhi, V ; Arora, V ; Behera, Laxmidhar ; Prasad, G ; Coyle, DH ; McGinnity, TM. / EEG denoising with a recurrent quantum neural network for a brain-computer interface. Unknown Host Publication. 2011. pp. 1583-1590
@inproceedings{17158662c07442049ba6337e3b1d0c6d,
title = "EEG denoising with a recurrent quantum neural network for a brain-computer interface",
abstract = "Brain-computer interface (BCI) technology is a means of communication that allows individuals with severe movement disability to communicate with external assistive devices using the electroencephalogram (EEG) or other brain signals. This paper presents an alternative neural information processing architecture using the Schrödinger wave equation (SWE) for enhancement of the raw EEG signal. The raw EEG signal obtained during the motor imagery (MI) of a BCI user is intrinsically embedded with non-Gaussian noise while the actual signal is still a mystery. The proposed work in the field of recurrent quantum neural network (RQNN) is designed to filter such non-Gaussian noise using an unsupervised learning scheme without making any assumption about the signal type. The proposed learning architecture has been modified to do away with the Hebbian learning associated with the existing RQNN architecture as this learning scheme was found to be unstable for complex signals such as EEG. Besides, this the soliton behaviour of the non-linear SWE was not properly preserved in the existing scheme. The unsupervised learning algorithm proposed in this paper is able to efficiently capture the statistical behaviour of the input signal while making the algorithm robust to parametric sensitivity. This denoised EEG signal is then fed as an input to the feature extractor to obtain the Hjorth features. These features are then used to train a Linear Discriminant Analysis (LDA) classifier. It is shown that the accuracy of the classifier output over the training and the evaluation datasets using the filtered EEG is much higher compared to that using the raw EEG signal. The improvement in classification accuracy computed over nine subjects is found to be statistically significant.",
author = "V Gandhi and V Arora and Laxmidhar Behera and G Prasad and DH Coyle and TM McGinnity",
year = "2011",
month = "7",
day = "31",
doi = "10.1109/IJCNN.2011.6033413",
language = "English",
pages = "1583--1590",
booktitle = "Unknown Host Publication",

}

Gandhi, V, Arora, V, Behera, L, Prasad, G, Coyle, DH & McGinnity, TM 2011, EEG denoising with a recurrent quantum neural network for a brain-computer interface. in Unknown Host Publication. pp. 1583-1590, The 2011 International Joint Conference on Neural Networks, 31/07/11. https://doi.org/10.1109/IJCNN.2011.6033413

EEG denoising with a recurrent quantum neural network for a brain-computer interface. / Gandhi, V; Arora, V; Behera, Laxmidhar; Prasad, G; Coyle, DH; McGinnity, TM.

Unknown Host Publication. 2011. p. 1583-1590.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - EEG denoising with a recurrent quantum neural network for a brain-computer interface

AU - Gandhi, V

AU - Arora, V

AU - Behera, Laxmidhar

AU - Prasad, G

AU - Coyle, DH

AU - McGinnity, TM

PY - 2011/7/31

Y1 - 2011/7/31

N2 - Brain-computer interface (BCI) technology is a means of communication that allows individuals with severe movement disability to communicate with external assistive devices using the electroencephalogram (EEG) or other brain signals. This paper presents an alternative neural information processing architecture using the Schrödinger wave equation (SWE) for enhancement of the raw EEG signal. The raw EEG signal obtained during the motor imagery (MI) of a BCI user is intrinsically embedded with non-Gaussian noise while the actual signal is still a mystery. The proposed work in the field of recurrent quantum neural network (RQNN) is designed to filter such non-Gaussian noise using an unsupervised learning scheme without making any assumption about the signal type. The proposed learning architecture has been modified to do away with the Hebbian learning associated with the existing RQNN architecture as this learning scheme was found to be unstable for complex signals such as EEG. Besides, this the soliton behaviour of the non-linear SWE was not properly preserved in the existing scheme. The unsupervised learning algorithm proposed in this paper is able to efficiently capture the statistical behaviour of the input signal while making the algorithm robust to parametric sensitivity. This denoised EEG signal is then fed as an input to the feature extractor to obtain the Hjorth features. These features are then used to train a Linear Discriminant Analysis (LDA) classifier. It is shown that the accuracy of the classifier output over the training and the evaluation datasets using the filtered EEG is much higher compared to that using the raw EEG signal. The improvement in classification accuracy computed over nine subjects is found to be statistically significant.

AB - Brain-computer interface (BCI) technology is a means of communication that allows individuals with severe movement disability to communicate with external assistive devices using the electroencephalogram (EEG) or other brain signals. This paper presents an alternative neural information processing architecture using the Schrödinger wave equation (SWE) for enhancement of the raw EEG signal. The raw EEG signal obtained during the motor imagery (MI) of a BCI user is intrinsically embedded with non-Gaussian noise while the actual signal is still a mystery. The proposed work in the field of recurrent quantum neural network (RQNN) is designed to filter such non-Gaussian noise using an unsupervised learning scheme without making any assumption about the signal type. The proposed learning architecture has been modified to do away with the Hebbian learning associated with the existing RQNN architecture as this learning scheme was found to be unstable for complex signals such as EEG. Besides, this the soliton behaviour of the non-linear SWE was not properly preserved in the existing scheme. The unsupervised learning algorithm proposed in this paper is able to efficiently capture the statistical behaviour of the input signal while making the algorithm robust to parametric sensitivity. This denoised EEG signal is then fed as an input to the feature extractor to obtain the Hjorth features. These features are then used to train a Linear Discriminant Analysis (LDA) classifier. It is shown that the accuracy of the classifier output over the training and the evaluation datasets using the filtered EEG is much higher compared to that using the raw EEG signal. The improvement in classification accuracy computed over nine subjects is found to be statistically significant.

U2 - 10.1109/IJCNN.2011.6033413

DO - 10.1109/IJCNN.2011.6033413

M3 - Conference contribution

SP - 1583

EP - 1590

BT - Unknown Host Publication

ER -